Abstract Background and Significance: Sexual violence perpetration is robustly associated with social contexts that support high levels of alcohol consumption. Beginning in 2007, the Arizona Department of Health Services (ADHS) developed and field-tested the Safer Bars curriculum with support from the CDC Rape Prevention and Education (RPE) funding program. Safer Bars is a bystander prevention training for staff of alcohol serving establishments. This application proposes a theory-driven, quasi-experimental effectiveness evaluation by an independent team at The University of Arizona. Our approach is significant because few evaluations have measured behavioral outcomes as well as the mediators that explain the observed changes in bystander behavior, or community-level outcomes associated with such interventions. Specific Aims: the specific aims are to: (1) Test a hypothesized mediational model of individual-level behavioral outcomes based on the Theory of Planned Behavior; (2) assess the effect of the program on bar-level risk management safety policies; and (3) examine the community-level impact of the Safer Bars program on reducing the number of police dispatches in areas with high densities of alcohol-serving establishments. Methods: A cluster-randomized trial design with crossover to randomize participants at the bar level into intervention and delayed intervention arms is proposed. Assessments are at baseline, completion of training, and at three and six months post-training. The sample bars will be selected using Geographical Information System (GIS) mapping of ?hot spots? for police dispatches within a three-mile radius of the three major Arizona universities, all of which have a high density of alcohol-serving establishments adjacent to campus. ADHS will focus their program delivery on bars within these identified geographic parameters. Our analytical model will assess a hypothesized moderated mediational model based in the Theory of Planned Behavior (TPB). Outcomes of interest include: (a) the extent to which Behavioral Intent, the central construct of TPB, mediates bar staff performance of intended bystander behaviors at three-month follow-up and the extent to which this relationship is moderated by personal characteristics and bar environment; (b) the effects of the program on bar-level risk management patron safety policies; and, (c) the community-level effects as demonstrated by number of police dispatches in areas with a high density of alcohol-serving establishments at baseline and at the end of the project using GIS data. With evidence of efficacy, Safer Bars is feasible for scale-up without excessive financial burden in other states, which all receive RPE funding.